I Miss RSS

Many my age may remember a set of tools built on RSS that were prevalent in the open web, but have increasingly waned in popularity. Some, like Google Reader and Yahoo Pipes, have ceased to exist. I still think a lot about Yahoo Pipes as one of the early (first?) no-code tools that could be used with RSS to filter various items out of a feed so a user had more control over what they were consuming. These were promising tools that excited me a lot about what the internet had to offer in terms of both access to information and empowering users with control of their own information flow. Though I still use RSS as a primary means of consuming information I miss the era of more tools and services being built around it.

Early in the Web 2.0 era, products like Tumblr were built with support for RSS. If you weren’t a Tumblr user or didn’t want to consume blogs in your Tumblr account dashboard, you could still subscribe to blogs via RSS (I did, and still do interact with several Tumblrs this way).

There are still some interesting products being built around RSS today. The podcasting ecosystem has largely grown up built on RSS. But these days, even in podcasts, companies like Spotify are moving away from building on RSS as a standard and towards their own proprietary methods of publishing and subscribing to content. Power has coalesced around recommendation algorithms.

Why has RSS languished?

There are a few reasons I think RSS had trouble surviving in web 2.0 era.

  1. RSS was hard to monetize effectively. One reason podcasts have survived so long built on RSS is they commonly monetize by in-podcast ads and that approach is compatible with using RSS (vs a targeted ad that is personalized and programmatically delivered to a specific user). This type of monetization is replicated in other RSS publications like Daring Fireball which still monetize via promoted posts that are delivered via RSS.
  2. RSS made it hard for publishers to really understand their users’ behavior and interaction patterns via RSS feeds. You see some of this manifested in the lack of analytics and tooling in the podcasting ecosystem, and that this is part of the value proposition Spotify and Apple are trying to communicate to podcasters who enter their respective ecosystems.
  3. Quality of personalization mattered more than control of personalization. Generating a personalized feed for an individual user would require a truly unique RSS feed (even if the user never went back to interact with the feed). This was probably pretty untenable for most publishers to take on the work of doing on the server side (from a cost and complexity perspective), especially without monetization or analytics. As a result, most publishers opened their whole feed or allowed simple keyword sorting and left any sort of personalization logic to tools like Yahoo Pipes or other end-user/client applications. Publishing personalized feeds ultimately became centralized (to align with the business model around that publishing - ads in most cases) and algorithms like FB’s, Twitter’s and more gained popularity among users, moving control over personalization out of their control.

The net result of all this was an average user experience that (compared to Web 2.0 social networks) was more complicated than clicking a “follow” button and letting someone else do the ranking and personalization. It required more complex knowledge of underlying standard and how it works in order to build something truly personalized and “custom” for a user.

What can we learn from RSS for the next gen of web experiences?

The promise of these more open and standardized technologies still remains - that users might get more control of the information they consume and be able to craft their own experiences on top. As no-code continues to proliferate and consumer awareness around control of data increases, I think the time is right for a new set of web tools built on standards like RSS. I say “like RSS” because I am fairly certain there are new approaches here (vs resurrecting old technologies) that would yield more successful results in today’s world - things that take into account the shortcomings I outline above.

There are a host of new crypto-centric technologies that could be used for something like this that could enable content syndication alongside better access control, personalization (client side and privacy preserving!) payment mechanisms, and more. If you’re building this I’d love to hear what you think!